After the Poll Wars—Explaining the Upset
Numerous theories are being tossed about among pollsters, political theoreticians and confounded laypeople alike in an attempt to explain the large discrepancy between the results of professionally conducted pre-electoral polls and the actual election. In the final month of the campaign, three internationally respected polling agencies showed the FSLN with a strong advantage over UNO. A Washington Post/ABC News poll conducted in February showed the FSLN with a 16-point lead. Univision’s second Nicaragua poll, carried out at the end of January, showed an 18-point FSLN lead. Greenberg-Lake’s final poll, conducted in mid-January, showed a 27-point FSLN lead. The field work for the three polls was conducted by independent Nicaraguan research firms—the Univision poll by Logos, the other two by Itztani. A third national firm, ECO, conducted two polls in January and February for the Central American University (UCA) that showed, respectively, a 36- and a 32-point lead for the Sandinistas.
The polls wrongly predicted the outcome of the election either because they were poorly done, because people didn’t tell the truth or because something very significant changed people’s minds between the close of the campaign and election day. Different pollsters maintain different theories. Two days after the elections, an article in La Prensa called all the above polls “Sandinista polls” and claimed that they “were invented with the goal of hiding reality,” pointing out that the polls sponsored by La Prensa and the National Endowment for Democracy correctly predicted Violeta Chamorro as the winner. But given the reputation of the former polling agencies and the willingness of local firms to make available all methodological and technical information, La Prensa’s accusation is unlikely. Even if these firms were truly “Sandinista,” “hiding reality” would only have hurt the FSLN since public opinion polls played an important role in the development of Sandinista campaign strategy. Advance notice that they might actually lose the election would more likely have worked to their advantage.
Both the pollsters themselves and international experts consider the methodologies used in the aforementioned polls to be technically sound. This leads to the most widely held assumption—that the people interviewed chose not to tell the interviewer who they were really going to vote for, even when asked to fill out a secret ballot. The next question, then, is why. According to one conspiracy theory, a significant number of voters wanted to avoid an FSLN victory at all costs, so they decided to keep their vote for UNO secret, believing that, if the FSLN suspected it might lose, it would change its campaign strategy and win more votes. This suggests an unlikely degree of sophistication in a country with little electoral experience like Nicaragua. In addition, the atmosphere throughout Managua on February 26—the real shock and state of mourning and absence of celebration—dispute this theory.
The most amusing theory of all is that of former US President and still rabidly anti-Sandinista Ronald Reagan—that UNO’s victory is actually a Sandinista conspiracy that will result in the demobilization of the contras and force UNO to undertake the arduous task of rebuilding Nicaragua’s beaten-down economy.
Fear of Repression? The most popular explanation in Washington, though reinforced by at least one Nicaraguan polling agency, is fear. Die-hard anti-Sandinistas think they have found long-sought proof that the FSLN runs a repressive totalitarian regime: respondents feared they would be victims of Sandinista repression if the government discovered that they sympathized with the opposition.
Government repression is inconsistent with contemporary Nicaraguan reality. Isolated incidents of abuse by some authorities have been documented, but there are no death squads or systematic, government-sponsored human rights violations as in El Salvador or Guatemala, where pre-electoral polls (including polls by Univision) have been consistent with election day results. Howard Schuman, director of the Survey Research Center at the University of Michigan and member of the International Commission on Pre-Electoral Polls in Nicaragua, suggests in a March 7 New York Times op-ed that perhaps the reason for the consistency in those countries is not that citizens trust pollsters more than in Nicaragua but that they distrust both pollsters and election day equally. At least in Nicaragua, says Schuman, people trusted that their vote would be secret on election day.
Cirilo Otero of Itztani supports the fear argument, explaining that when interviewed by pollsters, people “feared that their political preference would be discovered, and they’d become victims of Sandinista reprisals.” He calls their vote for UNO on election day—by secret ballot—“a vote of vengeance,” responding to that fear. But before election day, Itztani firmly maintained that respondents answered openly and honestly with no apparent fear or distrust.
CID-Gallup, with its ties to the US Information Agency and dubious reputation, might be expected to find fear even where it didn’t exist. Yet it reports, “CID found the Nicaraguan people in general to be most responsive to the polling process and displayed no fear in any part of the country to express an opinion.” Even Otero’s other comments don’t seem to be consistent with the image of a repressive Sandinista regime. He writes that the people “are not against the revolution as a whole” and their decision “to vote for a change [doesn’t] mean that their real choice is the National Opposition Union (UNO).”
Putting the blame on fear might be to Itztani’s advantage as a polling agency. To a certain extent, it frees it of responsibility for the inconsistency of' its results with the election results, which must concern its staff not only as professional pollsters but also because their polls may be partly responsible for the FSLN’s defeat. Also, by suddenly attacking the Sandinistas for being “repressive,” Itztani may be attempting to regain its own legitimacy as a polling agency in the eyes of those who have accused it’s pollsters of being Sandinista sympathizers.
Social Pressure—Do as thy Neighbors DoIn Nicaragua, there is certainly social pressure to support the revolution. Though polling interviews are supposed to be conducted in private, there is rarely such a thing as privacy in a Nicaraguan household; even neighbors can often hear regular conversation without effort. It is reasonable to expect that some people who fully intended to vote for UNO decided to “play it safe” when responding to the pollster’s questionnaire—and everyone knows it is safer to go with the winner (sure to be the FSLN). This kind of social pressure is not that different from any society where certain behavior or attitudes are considered more acceptable than others. For some it might simply mean saying you support what appears to be the status quo. For others it might include an element of shame or guilt over choosing personal gain over national pride and resisting US pressure.
Even in the United States, citizens sometimes hide their real positions from pollsters for the same kinds of reasons. In the New York City mayoral race last fall, numerous polls showed David Dinkins, a black Democrat, with a strong lead over Republican rival Rudolph Giuliani. A New York-Newsday poll released five days before the election showed Dinkins with 50% of the vote as compared to Giuliani’s 39%. In a contest where everyone agreed that “race is not the issue,” Dinkins won, but by a very small margin. So, what was wrong with the polls? In the final analysis, a large percentage of Democrats interviewed were unwilling to admit, even to a stranger in private, that their own racism would drive them to vote for a white Republican over a black Democrat.
The “Güeqüense Factor” There is no history of civic culture, polling or even peaceful transfers of power in Nicaragua. A dictatorship that spanned several decades was ousted less than 11 years ago, leaving a deeply ingrained distrust of authority in the population. Another analysis of the poll results thus leads to the cultural tradition of the güeqüense. In discussing the effect of Nicaraguans’ social psychology on public opinion polling, Logos director Marvin Saballos writes, “Distrust toward the dominant power appears to be profoundly rooted in the collective memory of the Nicaraguan people, who try to diminish that power by making good use of the methods of satire and irony.” “El Güeqüense” is an anonymous comedy ballet from the 17th century, written in old Spanish and corrupted Nahualt, an indigenous language. It is still often performed in the indigenous communities of Nicaragua.
“The Güeqüense,” continues Saballos, “the people’s representative, makes fun of the representative of authority, confusing him with a play on words so that the colonial authority never knows what the people really think” and leaves believing that they agreed with his point of view. Saballos asserts that the Nicaraguan people “güeqüensed” the pollsters. Associating certain polling agencies and interviewers with “the authorities” led them to respond in the manner in which they believed the interviewer wanted them to respond. Saballos claims that ECO and Itztani were associated with the FSLN, while La Prensa’s polls were clearly associated with the UNO. A certain sector of those interviewed, he says, trusted the latter more than the “representative of authority.”
Logos” strategy was to maintain a low profile in order not to be associated with either side. Its first poll, which was criticized for technical reasons (see “Poll Wars,” February envío), showed the FSLN in the lead over UNO by less than 1%: 40 to 39.1. By the time Logos conducted its second poll, says Saballos, voters were more cautious for three main reasons—the growing intensity of the “poll wars,” the tense military climate on the heels of the Panama invasion and the proximity of election day—distorting the results in favor of the FSLN.
Unlike other pollsters who have scrambled to understand the election results, Saballos maintained this theory throughout the pre-electoral period, suspecting all along that certain sectors of the population were responding according to the political leaning they attributed to the interviewer. In the United States, market researchers are fully aware of such dynamics, related to güeqüense psychology or to social pressure to conform, and systematically design their polls to compensate. Adjustments of this nature need to be made in future polls conducted in Nicaragua.
A Last Minute Change of Heart? Though perhaps wishful thinking, it is also possible that the polls were not wrong at all. After conducting a preliminary post-election study interviewing people who voted for UNO, ECO maintains that if the election had been held one or two weeks earlier, the results would have been much closer to that of its final poll, which showed a sweeping FSLN victory. ECO’s polls were exactly accurate in calculating the amount of solid FSLN support at 41% of the voting population, which would mean, according to their hypothesis, that all of the not so-solid voters decided at the last minute to vote for UNO. What could have swayed these people toward UNO in the week between the last poll and election day?
Random interviews in the days following the elections brought out the military draft as one of the primary issues on people’s minds. ECO’s interviews in areas where UNO won by a solid margin led it to the conclusion—which it hopes to test through further study—that, through a combination of circumstances election day became a referendum on the military draft instead of a vote for President. Oscar Medrano of ECO explains that the “undecided” population (referring to the 40% ECO considers neither solid UNO nor solid FSLN) was contemplating several controversial issues up until election day, and in the end chose the lesser of two evils. On one end of the scale were the draft and the contras; on the other, UNO. The FSLN meant a continuation of the draft; UNO meant an end to the draft. The contras were the reason the draft existed, and UNO meant the end of the contras. Ipso facto, UNO became the lesser evil, explains Medrano.
“So why didn't this happen until February 25, if on the 16th they told us they were going to vote for the FSLN?” Medrano asks rhetorically. Because, he continues, several Sandinista leaders said that “spectacular announcements” would be made at the FSLN’s closing campaign rally on February 21. A popular rumor spread that President Ortega was going to announce the end of the draft. ECO determined that “a large number of undecided voters, who had never been to a demonstration before, went to that rally because they wanted to hear something that in fact wasn’t said,” says Medrano. On the contrary, Ortega reaffirmed the need for military service.
Expectations were high, explains Medrano, and people did not hear what they wanted to hear. On election day they issued a “punishment vote” against the draft, fully believing the FSLN would win.
César Jérez, rector of the Central American University that sponsored ECO’s pre-election polls, does not share ECO’s analysis. Jérez says, “We simply have to confess that the polls did not reflect the voters” intention to vote and that we didn’t use scientific control instruments to balance the interpretations we were making of the polis. We were wrong. The people taught us a lesson.”
A major concern arises over ECO's new post-election study—if the güeqüense theory is correct, how does ECO know that respondents are not still answering questions the way they believe ECO wants them to respond? Medrano says that the post-election study shows a clear pattern that fit its previous polls’ calculations of solid FSLN, solid UNO and “undecideds.” When asked if an end to the draft would have affected the election results, the solid FSLN supporters said that the FSLN would have won by a large margin, the undecideds said they would have voted for the FSLN, and the solid UNO supporters said, “Nobody would have believed it.”
ECO's explanation is appealing, and there is undoubtedly some truth to it. Reporters found numerous people on February 26 who mentioned the draft as a primary issue and the February 21 rally as a real disappointment. At least some people who voted for UNO were repentant, shocked that the FSLN had lost. There was almost no celebration of the UNO victory; the mood in Managua was that of a massive funeral. But just how many people were repentant, how many people cast a “punishment vote” for that reason may be impossible to ever measure.
The UNO Polls—Better, or Just Lucky? The fact that other polling agencies came closer to the actual election results does not necessarily guarantee that their polls were done more accurately or even that they were done accurately at all. For example, the sponsor of CID-Gallup’s second poll reportedly chose not to release the results, presumably strongly in favor of UNO, because the methodology and field work were so outrageously unprofessional that he risked significant damage to his reputation. Some pollsters believe that people interviewed felt more comfortable revealing their true leanings, toward UNO, to interviewers considered pro-UNO. But prior criticisms of those firms’ methodologies and questionnaires still hold true (see February envío).
Borge and Associates, who conducted four polls for La Prensa, remains an object of controversy. Some pollsters, like Otero, congratulate the firm for its accuracy, while pointing out that it is still impossible to do a real analysis without the polls’ technical information. Others, like Medrano, call Borge’s Nicaragua polling “a charade,” saying he “won the lottery without buying a ticket.” A person working closely with La Prensa in the conduct of Borge’s polls privately criticized the training of interviewers and the numerous problems this caused in the field, among other aspects of the polling.
The pollsters from the firms of ECO, Itztani and Logos are in general agreement that their polls are methodologically and technically sound. Itztani believes that the only questions significantly affected by deceptive answers were those relating directly to voter preference, and, therefore, that their polls are still reliable tools with which to measure public opinion. Otero reports that a closer analysis of the answers to questions about military service and Sandinista economic management, for example, reveals tendencies that would have suggested a lean toward UNO. Itztani’s mistake, says Otero, was to give too much weight to the questions about voting and not conduct an in-depth analysis and cross tabulation of other responses. Otero points out that even Greenberg-Lake, which did extensive cross-tabulations, didn’t fully analyze the results.
The changes Itztani intends to make in future polls include careful analysis of the tone and content of each question, avoiding phrases or word choices that indicate a lean to either the right or the left, and improved selection and training of interviewers to assure that their political positions do not interfere with the polling process. Itztani’s changes appear to respond more to the güeqüense theory than to their own “fear” theory.
Marvin Saballos believes that responses to the questionnaire as a whole were affected by the güeqüense psychology. He suggests that some questions were probably answered more spontaneously than others, such as those referring to the economic situation or the country's problems in general. Future polls, says Saballos, should be accompanied by other kinds of research that provide additional “in-depth qualitative information that aid in the interpretation of the poll results.” Lengthier interviews should be conducted with the population to try to determine the respondent's real thoughts and assure that he or she isn’t simply spouting a certain political line, says Saballos, This would demand better training of interviewers. He argues that polls on subjects that are not as sensitive as this election or at times when the country is not as polarized can still be valuable tools with which to study public opinion in Nicaragua.
It would be of great value to the future of polling in Nicaragua if Borge and Associates’, in accordance with international ethical standards, were to disclose its methodology and questionnaires. Until then, it will not be possible to analyze its Nicaragua polls fully. If, in fact, the firm’s methodology does meet professional standards, its polls could provide important insights into the current debate, like a key missing piece to a very confusing puzzle. Before election day, of course, a poll certified to be technically sound that showed UNO in the lead might have provided a timely warning to the FSLN.
A Vote For UNO, A Vote Against the FSLNThe theories explored by the different pollsters focus on understanding the discrepancy between the polls and the election results, not specifically on why 55% of the population voted for UNO. All the polls do consistently point to two issues as the Nicaraguan people’s primary concerns—the economy and the war, fully describes in this month’s lead article. Analysts appear to have misread which party was believed best able to resolve them; at least in some cases, this does indicate a lean toward UNO. For example, in Greenberg-Lake’s January poll, although many people also blame the US for the economic situation, 46% agreed to the statement that the FSLN was “responsible for the economic decline of the country”; 44% disagreed. While direct questions implied that people look to the FSLN as more capable of bringing peace to Nicaragua, this issue is clearly affected by relations with the US. Of those polled, 49% agreed with the statement that the FSLN is “too hostile to the US.” On the other hand, 61% (compared to 5O% for the FSLN) believed that UNO would be ”able to reconcile Nicaragua and the US,” and 58% (versus 25% for the FSLN) believed UNO would end conscription.
There is reason to believe that the Nicaraguan people issued a “punishment vote” against the Sandinistas based on a variety of factors. The first is the excessive and unequal weight of the economic adjustment program, which fell on the backs of state employees, peasants, members of the economy’s urban informal sector, agricultural laborers and even industrial workers. A second factor may have been the government’s lack of clarity about what are now predominantly civilian functions of the military, and the governments indecision about transferring it from obligatory to voluntary service on condition that the war not reach new levels posing a strategic threat to the revolution. These two fundamental factors are brutal and efficient effects of the “low-intensity” war imposed on Nicaragua by two US administrations.
Other factors also influenced the election results to varying degrees. One is the dilemma between war and peace mentioned in the above polls, where the FSLN’s challenging tone to the US—in the context of national dignity—posed a threat to peace. Another is the breach that exists between the revolution’s fundamental goals—mixed economy, political pluralism, non-alignment, participatory democracy, respect for all ideologies—and its history of top-down, bureaucratic work styles, which still affect all levels of party militancy, party organizations, mass organizations and state institutions to a greater or lesser degree.
Scenes from the Day AfterReactions to the election results indicate that few suspected so many people would vote for UNO. On February 26, all of Nicaragua, not just the 41% that voted for the FSLN, was in mourning. UNO supporters did not pour into the streets to celebrate—there was almost no celebrating to be found. The streets were silent as people reacted with shock and uncertainty. President Ortega’s concession speech (reprinted in this issue) deeply moved a good number of Nicaraguans. It showed great respect for the Nicaraguan people without the least reproach. On the contrary, he exalted them perhaps more than ever before as a dignified and heroic people who have won the admiration and raised the hopes of all those who aspire to peace with justice for weak and impoverished peoples everywhere.
A major element in the public’s response to the election, and in the results themselves, was the overwhelming general belief that the FSLN would win. Not just the FSLN, but people throughout Nicaragua, the UNO leadership and even the US government were surprised by the upset. The polls, the size of the FSLN rallies—much larger than UNO’s—and the Sandinistas’ somewhat arrogant triumphalism that set the bold goal of beating their 67% victory in 1984 all convinced people that the FSLN was unbeatable. Had they not believed that, many who decided to vote for UNO would have voted for the FSLN. This is not just speculation.
The popular acceptance of a Sandinista victory may have had the uncalculated effect of encouraging many Nicaraguans to feel that their vote, by itself, was insignificant, because the FSLN did not need it to win. Some hypothesize that this may have led to a “punishment vote” against the FSLN, not to defeat it but to limit its margin of victory. Instead of the sweeping mandate of 1984, they preferred a strong opposition to impede the abuses of power that often occur when a party in power is unchallenged.
The firm belief that the FSLN would win helps explain some post-election scenes. Several municipal candidates in towns where UNO won now do not want to take office. After the results were in, Maria Luisa, who voted for UNO, ran to greet a friend sobbing, “We lost!” At the UNO victory rally in Santo Tomás, the incumbent mayor had to calm down the angry crowd by reminding them they had voted for UNO and should let the mayor-elect go home peacefully. Some mothers of fallen combatants, many of whom are part of one of the most patriotic and revolutionary organizations in the country, have sent letters to the Women’s Association office in León expressing regret at having voted for UNO, explaining that they feared losing another draft-age son.
Some people admitted they had cast a sympathy vote for Violeta Chamorro, widow of the martyred and mythologized La Prensa editor, ridiculed for her inability to give a political speech without reading it or to read it without errors, a nice lady with the pluck to stay on the campaign trail even with her leg in a cast. They explained sheepishly that, given the polls and the Sandinista triumphalism, they didn’t want her to lose so badly.
The contradictions abound. One older woman who voted for UNO, and did not regret it, was asked what should be done if the contras refuse to demobilize before April 25. She replied, “We should take up arms and be done with them.” For her, as for many others, the principal issue is the economy. Rumors in the market places run wild—that no one will have to pay taxes any more, that no one will have to pay electric or water bills, that bank loans taken out under the Sandinistas will not have to be paid back. On a city bus one recent morning, three young men boarded and refused to pay, saying that with Violeta as President, no one would have to pay for public transit any more.
But these bubbles of economic expectation burst rather quickly. Just two days after the election, a worker behind a store counter said, “I wonder how soon the price of gasoline will go down?” A co-worker turned to her and laughed, “Are you kidding? It's not going to go down; it’ll probably go up!” A look of real distress came over the first as she said, ”But if prices are going to go up, why did we vote for UNO?”